Sleepy at the wheel. Analysis of the extent and characteristics of sleepiness among Belgian car drivers. Motor vehicle crashes that are due to driver sleepiness are often particularly severe. They usually occur on monotonous high-speed roads and typically involve a drifting vehicle that hits an obstacle at full speed. According to different international estimates, about 20% of all severe road crashes may be attributed to sleepiness at the wheel. This share is in fact similar to that of driving under the influence of alcohol (25% according to the 2009 European SafetyNet project). Measuring the prevalence of drowsy driving and its role in crash causation is, however, not self-evident. The main reason is that reliable measurement protocols are not available. It is highly challenging to estimate sleepiness purely on the basis of physical characteristics (certainly after a crash occurred) and when drivers that caused a crash can be interviewed, there is a general bias not to report sleepiness – either by unawareness or by unwillingness. As a result, crash reports are usually not accurate with respect to driver sleepiness and few data exist on the overall prevalence of sleepiness at the wheel. In the vast majority of international studies, the estimation of this overall prevalence is limited to the occurrence of sleepiness at the wheel across a relatively broad time interval. For instance, in Belgium, the 2012 BRSI attitudes survey asked respondents: ”How frequently have you felt tired and sleepy while driving in the past year?”(Meesmann and Boets, 2014). It appeared that in 58% of the cases this occurred at least once. Although such a percentage is informative, it collapses driver sleepiness across all single trips that a driver has made during the last year. Hence, it does not deal with the critical question of how much driving involves sleepy drivers at a given point in time. The goal of the present study is twofold. The first aim is to obtain a trip-based estimate of the prevalence of sleepiness at the wheel among Belgian car drivers. Sleepiness is measured with respect to an actual driving episode (a trip from A to B) instead of a certain time interval (e.g., during the last 12 months). A trip-based prevalence estimation can be compared with what is done in road-side studies on driving under the influence of alcohol where drivers are stopped to measure their blood-alcohol concentration (e.g., Riguelle, 2014). The current study introduces a new methodology: via an online questionnaire drivers were asked to reflect on a single journey they made during the last 24 hours and to indicate the level of sleepiness they experienced while driving on the Karolinska Sleepiness Scale (KSS). This new methodology reduces the risk of response bias (not admitting sleepiness at the wheel) because it avoids direct contact between respondents and researchers (as in a road-side set-up, for instance). At the same time it is feasible to reach a large sample of drivers that is representative of the Belgian car drivers population. The second goal of the present study is to understand the prevalence of sleepiness at the wheel among Belgian car drivers through a wide range of contextual variables. Apart from acute sleepiness during the journey, drivers were also asked about trajectory features, sleep habits, driving behaviour and several socio-demographic variables. Chronic sleepiness was assessed via the commonly used Epworth Sleepiness Scale (ESS; Johns, 1991). BRSI organized a web-based survey between June 15th and July 15th 2014. Over 2,500 respondents, drawn from a panel of 130,000 individuals completed the survey. At the start of the survey, participants indicated whether they drove a car within the last 24 hours. Immediately afterwards, they were asked to bring one of the journeys to mind and answer questions about that journey as accurately as possible. The journey of interest was determined randomly. The survey was structured according to six topics (see Appendix 2 for the full survey): *Physical characteristics of the trajectory *Sleepiness during the journey *Last sleep episode before the journey *Driving behaviour *Fatigue *Socio-demographics The main dependent variable was sleepiness during the journey as measured on the Karolinska Sleepiness Scale. The results show that overall, 4.8% of the journeys by car drivers in Belgium involve a driver that is showing signs of sleepiness. Figure A shows the obtained distribution with respect to the separate levels of the Karolinska Sleepiness Scale. The analysis of contextual variables shows that various circumstances result in a prevalence that is considerably higher than the overall estimate of 4.8%. A regression analysis reveals unique effects of the following contextual variables on the prevalence of driver sleepiness (in decreasing order of effect sizes; prevalence estimates appear between brackets): 5. Spending more than 4 hours a day at the wheel (31%) 6. Having caught less than 8 hours of sleep (4-8 hours: 11% on average; 0-4 hours: 25% on average) 7. Having an irregular sleep-wake pattern with frequent shifts of more than 2 hours (15%) 8. Being an adolescent/young adult (18-30 years: 11% on average) 9. Having consumed 2 or more standard units of alcohol prior to driving (2-4 units: 11% on average) 10. Having caused a crash or a near-crash during the past 12 months (11%) 11. Experiencing excessive daytime sleepiness (9<ESS?15: 8% on average; ESS>15: 11% on average) 12. Long distance driving (>60 km: 11% on average) 13. Driving in the evening or at night (6-12pm: 8% on average; 12pm-6am: 8% on average) Figure B illustrates the continuous nature of these effects (with the exception of the binary crash history variable). Individual effects together with the 95% confidence bands are shown in blue. Irregular grey lines show the estimated prevalence based on the combination of all individual effects. The percentages in the bottom give the same prevalence, but aggregated into categories delimited by the vertical lines. Distributional analyses also show significant associations with the prevalence of driver sleepiness for the following categorical contextual variables. 14. Having a full-time job (8%) 15. Having a master’s degree (7%) 16. Being an employee (7%) or a manager (9%) 17. Dealing with circumstances with a chronic negative effect on sleep quality: stress/depression (8%), long lasting sleep interruptions (7%), obligation to get up early (11%), superficial sleep (8%), difficulties falling asleep (9%), irregular working hours (12%), excessive snoring (8%), family members with sleep problems (9%) and chronic insomnia (12%) This study reveals that, at the level of individual trips, on average 4.8% of the Belgian car drivers shows signs of sleepiness. Although there are no exact Belgian data on the role of sleepiness in crash causation, international numbers suggests that sleepiness at the wheel accounts for about 20% of all severe crashes. The combination of a relatively low prevalence and a relatively large share in crash causation implies a very important risk and shows similarities with driving under the influence of alcohol. According to the most recent estimates, 2.4% of all driving in Belgium occurs under the influence of alcohol (Riguelle, 2014) whereas the share in severe accident causation amounts up to 25% (SafetyNet, 2009). Hence, like driving under the influence of alcohol, the importance of sleepiness at the wheel for road safety should not be underestimated. This study clearly demonstrates that the prevalence of sleepiness at the wheel varies greatly with specific circumstances. The majority of these circumstances have been documented before, but it is the first time that they become quantified jointly in a Belgian context. Based on this quantification the scenario with the highest risk for driver sleepiness appears to be the following: A young person who caught less than 8 hours of sleep is driving a car for a long distance around midnight after having consumed some alcohol. He or she drives a car frequently and while doing so, caused a crash or near-crash in the past 12 months. He or she also has an irregular sleep-wake pattern and often feels sleepy during the day. The following recommendations can be made: Infrastructure: It is common practice in Belgium to implement rumble strips to alert drivers when their vehicle is drifting. Several studies have demonstrated a high benefit-to-cost ratio for this measure and further implementation can thus be encouraged. If a sleepy driver is alerted by rumble strips, this does not mean that he/she will stay alert for the rest of the trip, however. The creation of more safe(r) rest areas is an infrastructural investment that is more costly, but also more beneficial since the goal is to eliminate sleepiness (see Reyner et al., 2010). Apart from creating rest areas, signalling their presence to drivers is also an important point of action. Technology: Further development of in-car and wearable solutions for detecting sleepiness at the wheel needs to be encouraged. However, to attain maximal transparency about their abilities and limitations, current and future technologies need to be validated through independent research. Raising awareness: Campaigns are needed to inform drivers about the risks of drowsiness. The fact that sleepiness at the wheel is potentially as dangerous as driving under the influence of alcohol can help improving awareness. Perhaps more important than informing drivers about the risks of drowsiness, campaigns should focus on effective strategies to fight sleepiness at the wheel and to avoid it in the first place. It is important to acknowledge that suitable rest areas and/or other fit drivers are not always available when drivers start experiencing sleepiness. Hence, it is also critical to encourage drivers to plan their trips in advance. Sleep hygiene: Sleepiness at the wheel is part of a broader problem in today’s society, namely that of neglecting healthy sleep habits. It therefore needs to be approached from a broader perspective than road safety management alone. At the individual level people need to be informed about the health risks due to poor sleep habits (e.g., obesity, heart disease, diabetes and cancer) and the aspects of their daily lives that compromise a good sleep hygiene (e.g., exaggerated use of multimedia devices). Employers should also become aware about the impact of bad sleep habits among their employees. They can play an important role in reducing sleepiness and driver sleepiness in particular, for instance, by facilitating flexible work schedules and the flexible use of different transport modes for commuting. Further research: Investments are needed to further assess the impact of driver sleepiness on road safety. It is crucial to measure sleepiness at the wheel with respect to single journeys, in contrast to what has become common practice. The current method allows to obtain a trip-based prevalence on a large scale and in a relatively cost-effective manner. Provided sufficient resources are available, it is fairly straightforward to monitor the evolution of drowsy driving throughout the year and in different countries using the same method. Investigating the actual impact on road safety nevertheless also requires accurate numbers regarding crash causation. As in many other countries, accurate numbers on the role driver sleepiness in road crashes are lacking in Belgium, mainly because of the existing protocols in crash reporting and the lack of in-depth crash investigations. This is perhaps the area where research investments are needed most urgently. regulations. (Author/publisher)
Abstract